Previous disorders and depression outcomes in individuals with 12-month major depressive disorder in the World Mental Health surveys

Aims Major depressive disorder (MDD) is characterised by a recurrent course and high comorbidity rates. A lifespan perspective may therefore provide important information regarding health outcomes. The aim of the present study is to examine mental disorders that preceded 12-month MDD diagnosis and the impact of these disorders on depression outcomes. Methods Data came from 29 cross-sectional community epidemiological surveys of adults in 27 countries (n = 80 190). The Composite International Diagnostic Interview (CIDI) was used to assess 12-month MDD and lifetime DSM-IV disorders with onset prior to the respondent's age at interview. Disorders were grouped into depressive distress disorders, non-depressive distress disorders, fear disorders and externalising disorders. Depression outcomes included 12-month suicidality, days out of role and impairment in role functioning. Results Among respondents with 12-month MDD, 94.9% (s.e. = 0.4) had at least one prior disorder (including previous MDD), and 64.6% (s.e. = 0.9) had at least one prior, non-MDD disorder. Previous non-depressive distress, fear and externalising disorders, but not depressive distress disorders, predicted higher impairment (OR = 1.4–1.6) and suicidality (OR = 1.5–2.5), after adjustment for sociodemographic variables. Further adjustment for MDD characteristics weakened, but did not eliminate, these associations. Associations were largely driven by current comorbidities, but both remitted and current externalising disorders predicted suicidality among respondents with 12-month MDD. Conclusions These results illustrate the importance of careful psychiatric history taking regarding current anxiety disorders and lifetime externalising disorders in individuals with MDD.


Introduction
Depressive disorders are ranked as the third leading cause of 'years lived with disability' worldwide (GBD 2017 Disease andInjury Incidence andPrevalence Collaborators, 2018). This is due in part to their high prevalence, with a lifetime prevalence of 10.6% and a 12-month prevalence of 4.5% for major depressive disorder (MDD) (Bromet et al., 2018). Additionally, MDD is often characterised by a recurrent course (Hardeveld et al., 2013), suggesting the importance of taking a lifespan perspective when studying MDD (Hardeveld et al., 2010;Bockting et al., 2015).
Comorbidity with other mental disorders is high in individuals with MDD (e.g. Grant, 1995;Kessler et al., 2003;Alonso and Lépine, 2007;Bromet et al., 2018). Since MDD has a rather late median age of onset (AOO) (Kessler and Bromet, 2013), i.e. around 38 years (Bromet et al., 2018), other mental disorders with earlier AOOs, such as anxiety disorders (Goodwin, 2002;Bittner et al., 2004;Beesdo et al., 2007;Mathew et al., 2011;Kessler et al., 2012;Gundel et al., 2018) and behavioural disorders (including oppositional defiant disorder (ODD), conduct disorder (CD) and attention deficit hyperactivity disorder (ADHD)) (Kessler et al., 2012;Gundel et al., 2018) often precede first-onset MDD. Assessing the clinical history of mental disorders in individuals with current MDD could provide valuable information regarding management and treatment strategies, especially since this history may be a prognostic marker of the further clinical course of MDD (Wittchen et al., 2003). Mental disorder comorbidity is predictive of depression outcomes, including impairment (Andrews et al., 2001;Alonso et al., 2004;Kessler et al., 2015) and suicidal thoughts and behaviours (Stein et al., 2001;Bolton et al., 2010;Kessler et al., 2015). However, to our knowledge, no study has comprehensively examined the associations between a wide range of mental disorders with an onset before 12-month MDD diagnosis with multiple depression outcomes.
The overall aim of the present study is to examine associations between 12-month MDD, previous mental disorders and MDD outcomes. First, we examine the strength of associations between 12-month MDD and preceding individual lifetime mental disorders in the cross-national World Mental Health (WMH) surveys. Second, we examine whether previous mental disorders, i.e. with an onset prior to the respondent's age at interview, are predictive of current depression outcomes including impairment and suicidality.

WMH surveys
Data came from 29 cross-sectional surveys administered between 2001/2002 and 2015 in low/middle-income countries (Brazil, Bulgaria, Colombia, Colombia (Medellin), China (Shenzhen), Iraq, Lebanon, Mexico, Nigeria, Peru, Romania, South Africa and Ukraine) and high-income countries (Argentina, Australia, Belgium, France, Germany, Israel, Italy, Japan, the Netherlands, New Zealand, Northern Ireland, Poland, Portugal, Spain, Spain (Murcia), USA). A total of 145 990 respondents participated. Adults were selected based on multistage clustered area probability sampling designs designed to generate samples that were representative of the household populations in each country. Face-to-face interviews were conducted in respondent homes. The details of sampling methods in the different countries are described in detail elsewhere (Heeringa et al., 2008;Pennell et al., 2008). The weighted average response rate across surveys was 69.5%. Informed consent was obtained according to protocols endorsed by local Institutional Review Boards. Further information on the surveys is provided in online Supplementary Table 1.

MDD and other comorbid DSM-IV diagnoses
The WHO Composite International Diagnostic Interview (CIDI), a fully structured interview administered by trained lay interviewers, was used to assess DSM-IV mental disorders (Kessler and Üstün, 2004). Interviews were administered in two parts to diminish respondent burden. All respondents completed Part I, assessing MDD and other core mental disorders. Part II, assessing other disorders and disorder correlates, was administered to all respondents with any lifetime Part I diagnosis and a probability subsample of other respondents. Part II data were weighted to adjust for the undersampling of Part I non-cases. Analyses for this study were performed in the Part II sample (N = 80 190).
Since mental disorders are significantly correlated, which may be indicative of shared underlying vulnerability factors, a modelbased approach of comorbidity of mental disorders was used (Krueger and Markon, 2006). Grouping of DSM-IV disorders was based on the underlying structure of disorders found previously in the WMH surveys (de Jonge et al., 2018). Given our focus on MDD, we separated two groups of distress disorders: depressive distress disorders (MDD and dysthymia (hierarchical)) and non-depressive distress disorders (generalised anxiety 2 Annelieke M. Roest et al. disorder (GAD) and post-traumatic stress disorder (PTSD)). Fear disorders included agoraphobia (non-hierarchical), panic disorder, social anxiety disorder, specific phobia and (childhood or adult) separation anxiety disorder. Externalising disorders included intermittent explosive disorder, ADHD, CD, ODD, alcohol abuse, alcohol dependence, drug abuse and drug dependence. The AOO of disorders was assessed using special recall probes that have been shown to yield more plausible distributions of AOO of disorders than conventional recall questions (Knäuper et al., 1999). Lifetime comorbid disorders were taken into account when they had an onset prior to the respondent's age at interview, regardless of whether or not they persisted into the past year.

Role impairment
Role impairment due to MDD was assessed with a modified version of the Sheehan Disability Scale (SDS) (Leon et al., 1997). This scale assesses impairment in different domains, namely home management, ability to work, ability to form and maintain close relationships, and social life, during the month in the past year when MDD was most severe. Severe impairment was defined as a score ⩾7 on a response scale of 0-10 for each domain. Respondents were defined as having any severe impairment if they scored ⩾7 in at least one domain. Additionally, respondents were asked how many days in the past year they were unable to work or carry out their normal activities because of their depression, i.e. days out of role (Ormel et al., 2008).

Suicidal thoughts and behaviours
Respondents were asked whether they had ever seriously thought about suicide, and if so, whether they had ever made a suicide plan or attempted suicide. Respondents who reported these experiences were asked whether these experiences had also occurred in the past 12 months.

Statistical analysis
The actuarial method was used to calculate the median AOO of MDD in respondents with 12-month MDD. We used logistic regression to examine the association between 12-month MDD and other mental disorders (controlling for country of origin). If a specific disorder was not assessed in a particular country, that country was not included in that analysis; however, if a country assessed at least one of the disorders in a disorder group (e.g. fear disorders), the country was included in the analysis of that disorder group, and non-assessed disorders were assumed not to be present. We also used logistic regression to examine the association between previous mental disorder groups and depression outcomes (role impairment and suicidality). Linear regression was used for days out of role due to MDD in the past year, a continuous outcome. The models included all disorder groups simultaneously. We tested three models: model 1 only controlled for country of origin of the participant; model 2 additionally adjusted for age and sex; and model 3 additionally adjusted for educational status, marital status and employment status.
In post-hoc analyses, we added depression characteristics, namely AOO of MDD, lifetime history of chronic MDD (duration of longest episode >2 years) and severity of MDD in the worst month of the past year (according to the Quick Inventory of Depressive Symptomatology Self Report (QIDS-SR)) (Rush et al., 2000), to model 3. Furthermore, in a previous study, current, but not remitted anxiety disorders and dysthymia, predicted MDD episode duration (ten Have et al., 2017). As our main analyses include both remitted disorders and disorders that had onset prior to the current age but persisted into the past year, we distinguished remitted (i.e. no 12-month diagnosis) from current disorders in additional post-hoc analyses.

Sample characteristics
A full description of the sample's characteristics stratified by 12-month MDD status is presented in online Supplementary Table 2. The number of respondents who met the criteria for 12-month MDD was 6684. The (weighted) prevalence of 12-month MDD was 4.9% (standard error (S.E.) = 0.1). Of the respondents with 12-month MDD, 66.8% were female. In this group, the median age of MDD onset was 38 years (interquartile range = 23-54) and 34.4% had a lifetime history of chronic MDD.
Among respondents with 12-month MDD, those with a previous depressive distress disorder were less likely to be severely impaired due to MDD in home management (OR = 0.6; 95% CI 0.4-0.8; p = 0.001), ability to work (OR = 0.6; 95% CI 0.4-0.8; p < 0.001) and social life (OR = 0.6; 95% CI 0.4-0.8; p < 0.001) compared to respondents without a previous mental disorder (see model 1, Table 2). However, the association between history of depressive distress disorders and ability to form and maintain close relationships was not statistically significant (OR = 0.7; 95% CI 0.5-1.0; p = 0.061). Respondents with a previous nondepressive distress disorder (GAD or PTSD) or fear disorder had significantly increased odds (ORs ranging from 1.3 to 1.7) of impairment due to MDD in all roles, compared to respondents without previous mental disorders. Respondents with a previous externalising disorder also had an increased risk of impairment in ability to work, social life and relationships (ORs = 1.2-1.3), but not of impairment in home management (OR = 1.0; 95% CI 0.8-1.2; p = 0.96). Adjustment for sex and age in model 2, and additionally for education, marital status and employment in model 3, did not change these findings (see Table 2).

Days out of role
The mean number of days out of role in the past year due to MDD for participants with 12-month MDD was 36.4 (S.E. = 1.8). Compared to respondents with 12-month MDD without previous mental disorders, those with previous depressive distress disorders (B = 10.9; 95% CI 1.1-20.8; p = 0.030), non-depressive distress disorders (B = 16.6; 95% CI 8.4-24.8; p < 0.001) or fear disorders (B = 13.8; 95% CI 6.4-21.3; p < 0.001) had higher numbers of days out of role (online Supplementary Table 3). The number of days out of role was not significantly higher for respondents with previous externalising disorders (B = 5.1; 95% CI −2.5 to 12.6; p = 0.19). After adjustment for sex and age, the association for previous depressive distress disorders lost statistical significance (B = 9.3; 95% CI −0.3 to 18.8; p = 0.057). However, previous non-depressive distress disorders (B = 13.2; 95% CI 4.9-21.4; p = 0.002) and previous fear disorders (B = 14.1; 95% CI 6.8-21.4; p < 0.001) remained significant predictors of days out of role even in the fully adjusted model.

Suicidal thoughts and behaviours
Of all respondents with 12-month MDD, 13.2% (S.E. = 0.5) reported suicidal ideation in the past 12 months and 4.4% (S.E. = 0.3) and 2.6% (S.E. = 0.2), respectively, reported suicide plans and attempts over this period. Previous depressive distress disorders were not significantly associated with 12-month suicidal ideation, plan or attempt in any model (ORs = 0.9-1.3) (Table 3). Previous non-depressive distress disorders were associated with suicidal ideation, plan and attempts in respondents with 12-month MDD, but not all associations were statistically significant in each model (ORs = 1.3-1.7) (see Table 3). Previous fear disorders (ORs = 1.5-1.7) and previous externalising disorders (ORs = 1.9-2.9) were consistently associated with suicidal ideation, plan and attempt in all models.

Post-hoc analyses including MDD characteristics
When depression characteristics were added to the models, a higher age of MDD onset and 12-month MDD severity were associated with role impairment due to MDD (online Supplementary  Table 4). In this model, previous fear disorders were no longer associated with role impairment in any of the roles, except social life (OR = 1.2; 95% CI 1.0-1.5; p = 0.029). Associations between  previous non-depressive distress disorders and previous externalising disorders with role impairment remained statistically significant (ORs = 1.1-1.4), with the exception of the association between previous externalising disorders and work impairment (OR = 1.1; 95% CI 0.9-1.4; p = 0.30).
Only previous fear disorders were associated with a higher number of days out of role compared to respondents with 12-month MDD without previous mental disorders (B = 7.2; 95% CI 0.0-14.4; p = 0.049) after additional adjustment for depression characteristics. Lifetime MDD chronicity and 12-month MDD severity were strongly related to days out of role in this model (online Supplementary Table 5).
Additionally, previous non-depressive distress disorders were no longer statistically significantly related to suicidal thoughts and behaviours. However, previous fear and externalising disorders (ORs = 1.4-2.1), as well as MDD severity, were associated with suicidal ideation, plans and attempts (online Supplementary Table 6).

Discussion
To our knowledge, this is the first epidemiological study that comprehensively examined the history of mental disorders in individuals with current MDD and their impact on depression outcomes. The results of this study show the importance of applying a lifespan perspective when studying MDD. We found that 95% of respondents with 12-month MDD had a previous mental disorder. Many had suffered from previous MDD episodes, confirming the recurrent and/or chronic nature of MDD and demonstrating that first-onset cases are rare among people with a 12-month MDD episode. However, even after excluding previous MDD, 65% of respondents reported previous mental disorders. Compared to respondents without MDD, those with MDD had higher lifetime prevalence rates of all other individual mental disorders.

Previous disorders and depression outcomes
Importantly, previous mental disorders in respondents with MDD were associated with impairment and suicidality. Previous non-depressive distress disorders and fear disorders were most strongly associated with impairment. These results are consistent with other studies that reported high impairment rates in individuals with MDD and comorbid anxiety disorders (Wittchen et al., 2003;Van Der Werff et al., 2010;Kessler et al., 2015) and reinforce the clinical importance of anxiety symptoms in MDD (Gaspersz et al., 2018). Further, previous non-depressive distress, fear and externalising disorders were related to past-year suicidal ideation, plans and attempts in respondents with 12-month MDD. The strongest associations were found for previous externalising disorders, consistent with other epidemiological studies that found associations between externalising psychopathology and suicide attempts independent from internalising psychopathology (Verona et al., 2004;Hills et al., 2005Hills et al., , 2009. Surprisingly, previous depressive distress disorders were associated with less frequent severe role impairment. However, respondents with previous depressive distress disorders did have a higher number of days out of role compared to respondents with first-onset MDD, although this association did not remain statistically significant after adjustment for covariates. Individuals with recurrent or chronic MDD may be less likely to report severe impairment in different domains as a result of habituation or adaptation to depressive episodes (Kruijshaar et al., 2003). For example, it is possible that recurrent or chronic MDD has led to the individual having reduced activity in some roles (e.g. being under-employed); an ability to function adequately in a reduced role may or may not be interpreted as impairment. However, individuals with recurrent or chronic MDD may also have reduced insight regarding impairment. The finding that people with previous depression do report a higher number of days out of role, a more objective measure of impairment, argues that reduced insight may play some role, yet this is a post-hoc speculation. Previous research has found conflicting results, with some other studies reporting that recurrence of MDD was related to increased functional impairment and work disability (Merikangas et al., 1994;Rytsälä et al., 2005), while others also found lower reported impairment among those with recurrent MDD (Kruijshaar et al., 2003;Van Der Werff et al., 2010), although this latter finding lost statistical significance after adjustment for depression characteristics, including symptom severity (Van Der Werff et al., 2010). Interestingly, we found that a history of depressive distress disorders remained a protective factor for role impairment even after adjustment for MDD characteristics, including severity.
Somewhat surprisingly, a history of depressive disorder was not associated with 12-month suicidal ideation, plan or attempt. This suggests that people with a longer history of depression are not currently at higher risk of suicidal thoughts and behaviours than people with first onset of depression. Nevertheless, it is likely  Association between severe role impairment and previous depressive distress, non-depressive distress, fear or externalising disorders, among respondents with 12-month MDD. Previous disorders are divided into remitted and current disorders. Depressive distress disorders: major depressive disorder and dysthymia; non-depressive distress disorders: generalised anxiety disorder and post-traumatic stress disorder; fear disorders: agoraphobia, panic disorder, separation anxiety disorder, social anxiety disorder, specific phobia; externalising disorders: intermittent explosive disorder, attention deficit hyperactivity disorder, conduct disorder, oppositional defiant disorder, alcohol abuse, alcohol dependence, drug abuse and drug dependence. Analyses are adjusted for country, age, sex, education, marital status, employment status and MDD characteristics (age of onset, chronicity, severity).
that people with a longer history of depression do have a higher lifetime risk of suicidality as risk increases over the years since persons with (a history of) MDD remain vulnerable for suicidal ideation over time (de Beurs et al., 2019).

Lifespan perspective to MDD and other mental disorders
Applying a lifespan perspective to MDD appears to be of importance, since individuals with MDD are highly likely to have suffered from previous MDD and other mental disorders and these disorders predict current, and potentially also future, depression outcomes. Other epidemiological studies have also shown that comorbidity is associated with increased depressive episode duration (ten Have et al., 2017), lower likelihood of MDD remission (Kelly and Mezuk, 2017) and chronicity of MDD (Satyanarayana et al., 2009;Murphy and Byrne, 2012;ten Have et al., 2018). On the other hand, research findings have not been consistent. For example, Hoertel et al. (2015) concluded that patient accounts of previous mental disorders would probably not improve the prediction of future suicidality, because current comorbid mental disorders explained the associations of remitted mental disorders with suicide attempts. Our post-hoc analyses also showed that most of the associations between comorbid mental disorders and outcomes were explained by current, rather than remitted, disorders, but our results contrast with Hoertel et al.'s since we found that both current and remitted externalising disorders were associated with 12-month suicidal thoughts and behaviours. Another study also suggested residual effects of (comorbid) psychopathology, finding that individuals with previous internalising and externalising disorders still reported higher functional impairment even 12 months after remission of the disorder (Bijl and Ravelli, 2000). Potential benefits of taking a lifespan perspective when examining mental disorders are probably not restricted to MDD. Future studies could examine the role of previous disorders in onset and associated disability of other mental disorders with a relatively late median AOO such as panic disorder, or instead, focus on early-onset disorders as causes or markers of future psychopathology. For example, in another report of the WMH surveys, specific phobia in childhood was predictive of other internalising disorders as well as impairment throughout the life course (de Vries et al., 2019).

Comparison with clinical studies
The results of the current study are in line with clinical studies. Among outpatients with a primary diagnosis of current MDD, 91% had another lifetime mental disorder diagnosis (including previous MDD) (Brown et al., 2001). Another study in outpatients with MDD also showed high prevalence rates of lifetime and current comorbidity (Zimmerman et al., 2002;Zimmerman et al., 2008). However, these findings do raise the question of whether these high comorbidity rates represent an artefact of current diagnostic systems, for example, as a result of the increasing number of diagnoses in the DSM (Maj, 2005a), while different diagnoses may actually result from a single clinical entity (Maj, 2005b). Nevertheless, detecting the presence of comorbid psychiatric symptoms may be very relevant in clinical management (Mai, 2005a). Indeed, the importance of taking comorbidity into account was exemplified by Zimmerman et al. (2002); their study showed that the majority of patients with primary MDD and comorbidity wanted treatment to also focus on these comorbid current disorders, partially remitted disorders or symptoms not meeting full diagnostic criteria (Zimmerman et al., 2002). Unfortunately, underrecognition of comorbidity is likely to be common in routine clinical care when unstructured clinical evaluations are used and time is limited (Zimmerman et al., 2008).
The importance of a lifetime approach to MDD in clinical care has been advocated before, for example, by Bockting et al. who argued that knowledge on risk factors of MDD recurrence, such as number of previous MDD episodes, may guide treatment choices (Bockting et al., 2015). The current study suggests the importance of additionally taking other previous mental disorders into account to further enhance personalised treatment, especially since most associations between previous lifetime comorbid disorders and current MDD outcomes remained statistically significant even after adjustment for MDD characteristics.

Strengths and limitations
An important strength of the current study is the large, cross-national sample, providing the opportunity to examine the associations between MDD and histories of individual mental disorders. Another strength is the comprehensive assessment of multiple depression outcomes and post-hoc analyses that adjusted for MDD characteristics and separated remitted from current comorbid mental disorders. Yet this study also has several limitations. First, disorders were retrospectively assessed, so recall bias may be present, particularly with regard to retrospectivelyreported AOO (Johnson et al., 1997). Second, although we distinguished first-onset from non-first-onset MDD episodes, we did not distinguish recurrent from persistent or chronic 12-month episodes. However, post-hoc analyses were adjusted for lifetime presence of chronic MDD. We also did not consider the duration or severity of comorbid disorders, although it is likely that more long-lasting and more severe comorbid disorders are more strongly associated with MDD outcomes. Finally, we did not consider the potential influence of prior treatment for either MDD or comorbid disorders.

Conclusions
In this general population sample, almost all individuals with 12-month MDD had suffered from previous mental disorders. Among these individuals, previous non-depressive distress and fear disorders were particularly strongly associated with greater role impairment, confirming the clinical importance of anxiety in MDD. Previous externalising disorders were most strongly associated with suicidal ideation, plans and attempts. The risks of adverse depression outcomes associated with previous mental disorders were reduced in strength, but did not disappear, after adjustment for MDD characteristics. Risks were mostly conveyed through current, rather than remitted, previous mental disorders, except for the risk of impairment in social life and suicide attempts associated with remitted externalising disorders. These results illustrate the importance of careful psychiatric history taking regarding current anxiety disorders and lifetime externalising disorders in individuals with MDD.
Supplementary material. The supplementary material for this article can be found at https://doi.org/10.1017/S2045796021000573.
Data. Access to the cross-national World Mental Health (WMH) data is governed by the organisations funding and responsible for survey data collection in each country. These organisations made data available to the WMH consortium through restricted data sharing agreements that do not allow us to release the data to third parties. The exception is that the US data are available for secondary analysis via the Inter-University Consortium for Political and Social Research (ICPSR), http://www.icpsr.umich.edu/icpsrweb/ICPSR/series/00527.